A diabetic milieu increases ACE2 expression and cellular susceptibility to SARS-CoV-2 infections in human kidney organoids and patient cells

Abstract

It is not well understood why diabetic individuals are more prone to develop severe COVID-19. To this, we here established a human kidney organoid model promoting early hallmarks of diabetic kidney disease development. Upon SARS-CoV-2 infection, diabetic-like kidney organoids exhibited higher viral loads compared with their control counterparts. Genetic deletion of the angiotensin-converting enzyme 2 (ACE2) in kidney organoids under control or diabetic-like conditions prevented viral detection. Moreover, cells isolated from kidney biopsies from diabetic patients exhibited altered mitochondrial respiration and enhanced glycolysis, resulting in higher SARS-CoV-2 infections compared with non-diabetic cells. Conversely, the exposure of patient cells to dichloroacetate (DCA), an inhibitor of aerobic glycolysis, resulted in reduced SARS-CoV-2 infections. Our results provide insights into the identification of diabetic-induced metabolic programming in the kidney as a critical event increasing SARS-CoV-2 infection susceptibility, opening the door to the identification of new interventions in COVID-19 pathogenesis targeting energy metabolism.

Keywords: ACE2; COVID-19; SARS-CoV-2; angiotensin-converting enzyme 2; diabetes 2; human kidney organoids.

Graphical abstract

Figure 1

High oscillatory glucose conditions induce early hallmarks of the diabetic kidney disease in human kidney organoids (A) Experimental scheme for the generation of human kidney organoids from hPSCs. (B) Trichrome Masson staining of control or diabetic kidney organoids. Glomerular (^∗) and tubular (^∗∗) structures are shown. Scale bars, 250 and 100 μm (magnified views). (C) Corresponding quantification of collagen fibers (B). y axis represents integrated intensity. Data are mean ± SD of at least n = 5 independent experimental replicates per condition. ^∗p < 0.05, unpaired Student’s t test. (D) Representative immunofluorescence staining for COLLAGEN-I (green), LTL (gray), and DAPI (blue) in control or diabetic kidney organoids. Scale bars, 250 and 100 μm (magnified views). (E) mRNA expression levels of COL3A1 and COL4A1 in control or diabetic kidney organoids. Data are mean ± SD. n = 3 independent biological replicates from a pool of 12 organoids/group with two technical replicates each. ^∗∗p < 0.01, unpaired Student’s t test. (F) Representative immunofluorescence staining for FIBRONECTIN (green), E-CADHERIN (ECAD; red), LTL (gray), and DAPI (blue) in control or diabetic kidney organoids. Scale bars, 250 and 100 μm (magnified views). Yellow arrows highlight sites of fibronectin deposits. (G) Seahorse analysis in LTL⁺ cells isolated from control or diabetic kidney organoids. The oxygen consumption rate (OCR) data are normalized to total protein. Data are mean ± SD. n = 10 biological replicates/group. ^∗∗∗∗p < 0.0001, two-way ANOVA, followed by Bonferroni post-test. Basal respiration and spare respiratory capacity, cellular ATP production, and maximal respiration are shown as mean ± SD. n = 10 biological replicates/group. ^∗p < 0.05; ^∗∗p < 0.005, unpaired Student’s t test. See also Figures S1 and S2 and Data S1.

Figure 2

Diabetic conditions induce ACE2 expression in human kidney organoids and enhance SARS-CoV-2 infections (A) Representative bright-field images and hematoxylin and eosin staining of control or diabetic kidney organoids. Asterisks highlight podocyte-like cells (^∗) or tubular-like structures (^∗∗). Consecutive sections were stained for ACE2 (green), PODOCIN (red), and using Lotus Tetraglobus lectin (LTL, tubular cell marker, gray) and DAPI (blue). Scale bars, 250 and 100 μm (magnified views). (B) Quantification of ACE2 colocalization with LTL⁺ and WT1⁺ cells in control or diabetic kidney organoids. Data are mean ± SD. n = 13 (control) and n = 16 (diabetic) organoids per condition. ^∗∗∗∗p < 0.0001, one-way ANOVA, Tukey’s multiple comparisons test. (C) Quantification of ACE2 colocalization with LTL⁺ and CD31⁺ cells in control or diabetic kidney organoids. Data are mean ± SD. n = 3 (control) and n = 4 (diabetic) organoids per condition. ^∗∗∗∗p < 0.0001, one-way ANOVA, Tukey’s multiple comparisons test. (D) Quantification of ACE2 colocalization with LTL⁺ and MEIS⁺ cells in control or diabetic kidney organoids. Data are mean ± SD. n = 4 (control) and n = 4 (diabetic) organoids per condition. ^∗∗∗∗p < 0.0001, one-way ANOVA, Tukey’s multiple comparisons test. (E) Quantification of the area of ACE2⁺ cells in control or diabetic kidney organoids. Data are mean ± SD. n = 20 (control) and n = 24 (diabetic) organoids. ^∗p < 0.05, unpaired Student’s t test. (F) Quantification of the mean fluorescence intensity (MFI as arbitrary units [AU]) of ACE2⁺ cells in control or diabetic kidney organoids. Data are mean ± SD. n = 6 (control) and n = 6 (diabetic) organoids. ^∗∗∗∗p < 0.0001, unpaired Student’s t test. (G) Protein levels of ACE2 in control or diabetic kidney organoids are shown by western blot. β-actin was used as loading control. The correspondent quantification is shown. Data are mean ± SD. n = 3 independent biological replicates from a pool of 12 organoids/group; ^∗p < 0.05, unpaired Student’s t test. (H) mRNA expression levels of ACE2 in control or diabetic kidney organoids. Data are mean ± SD. n = 4 independent experimental replicates from a pool of 12 organoids/group with two technical replicates each. ^∗∗p < 0.005, unpaired Student’s t test. (I) mRNA expression levels of ACE2 in control or diabetic kidney organoids upon actinomycin D treatment. Data are mean ± SD. n = 2 independent biological replicates from a pool of 12 organoids/group with three technical replicates each. ns, no statistical significance; ^∗∗∗∗p < 0.0001, two-way ANOVA followed by Bonferroni post-test. (J) Experimental scheme for the infection of control or diabetic kidney organoids with SARS-CoV-2. (K) Immunofluorescence of control and diabetic kidney organoids at 1 dpi with SARS-CoV-2 for ACE2 (green), viral nuclear protein (NP, red), LTL (magenta), and DAPI (blue). Scale bars, 250 and 50 μm (magnified views). n = 2 organoids per condition. (L) Quantification of viral NP colocalization with LTL⁺, CD31⁺, and WT1⁺ cells in control or diabetic kidney organoids. Data are mean ± SD. n = 3 control and n = 3 diabetic organoids. ^∗∗∗∗p < 0.0001, one-way ANOVA, Tukey’s multiple comparisons test. (M) Quantification of the area of NP⁺ cells from images in (K). Data are mean ± SD. n = 5 control and n = 5 diabetic kidney organoids performing two technical replicates each. ^∗p < 0.05, unpaired Student’s t test. See also Figure S2 and Data S1.

Figure 3

Diabetic-induced metabolic programming in human kidney organoids enhances SARS-CoV-2 infection (A) Uniform manifold approximation and projection (UMAP) of control or diabetic kidney organoids at 1 dpi with SARS-CoV-2. Clusters are colored by annotated cell types. (B) UMAPs for SARS-CoV-2 expression in control or diabetic kidney organoids at 1 dpi. For SARS-CoV-2, expression is considered as undetectable for cells expressing <5 unique molecular identifiers (UMIs). Cells are colored based on expression level. The violin plots in the bottom panels represent ≥5 UMI expression levels for the different cell types indicated. (C) A hallmark GSEA was performed separately for control and diabetic conditions, comparing SARS-CoV-2-infected versus mock organoids. The ten gene sets per direction and sample with lowest adjusted p values (p < 0.05) are shown. Each column corresponds to one of the comparisons. Circles are coded by color (direction), size (NES), and transparency (−log₁₀(p value)). (D) Differentially expressed genes (DEGs) in the comparison of the SARS-CoV-2-infected against mock organoids in diabetic conditions considering only renal-like cell types. In the volcano plot, the x axis indicates log fold change (FC), and the y axis indicates statistical significance with the −log₁₀(p value). Genes with an adjusted p value < 0.05 are considered upregulated (red) if the logFC > 0.1 and downregulated (blue) if the logFC < −0.1. Non-DEGs are shown in gray. See also Figures S3 and S4.

Figure 4

SARS-CoV-2 infection in human kidney organoids depends on ACE2 (A) mRNA expression levels of ACE2 in mock or SARS-CoV-2-infected kidney organoids. Data are mean ± SD. n = 2 independent biological replicates from a pool of 12 organoids/group with two technical replicates each. ^∗p < 0.05, unpaired Student’s t test. (B) Immunofluorescence staining for ACE2 (green) and LTL (magenta) in WT and ACE2 KO kidney organoids. Scale bars, 50 μm. (C) Experimental scheme for the infection with SARS-CoV-2 of WT and ACE2 KO kidney organoids. Immunofluorescence of mock or SARS-CoV-2-infected WT and ACE2 KO kidney organoids at 3 dpi for ACE2 (green), viral nuclear protein (NP, red), LTL (gray), and DAPI (blue). Scale bars, 250 μm. (D) mRNA expression of SARS-CoV-2 and ACE2 in mock or SARS-CoV-2-infected WT and ACE2 KO kidney organoids at 3 dpi by qPCR. Data are mean ± SD. n = 2 independent biological replicates from a pool of 12 organoids/group with two technical replicates each. ^∗∗∗∗p < 0.0001; ns, no statistical significance; one-way ANOVA, Tukey’s multiple comparisons test. (E) TEM analysis of WT and ACE2 KO kidney organoids infected with SARS-CoV-2 and recovered at 3 dpi. Representative images of infected ACE2 WT specimen show numerous viral particles (asterisks) inside a vesicle near the plasma membrane of a dying cell (1) and in the apical microvilli (amv) of a tubular-like cell (2). Details for podocyte-like cells exhibiting podocyte-related structures including primary processes (pp) (3), the deposition of a basement membrane (bm) (3), and apical microvilli (amv) (4) are shown. Scale bars, 10 and 5 μm, 200 nm (magnified views in 1 and 2), and 2 μm (magnified views in 3 and 4). Representative images of infected ACE2 KO specimen show epithelial tubular-like cells with brush borders (bb) (5) and tight junctions (tj) (6). Details for podocyte-like cells with podocyte-related structures including primary (pp), the deposition of a basement membrane (bm), and cell processes (sp) are shown. Scale bars, 5 μm, 500 nm (magnified views in 5 and 6), and 2 μm (magnified views in 7 and 8). See also Figures S5–S7.

Figure 5

SARS-CoV-2 infections in BSG KO kidney organoids (A) Schematic of Cas9/gRNA-targeting sites (pink arrows) in BSG locus showing exon structure (blue boxes) and PCR amplicons (light gray boxes in all figures in this study). Histogram shows allelic sequence distribution after the transfection of the different gRNAs in undifferentiated ES[4] cells expressing an inducible Cas9 (iCas9). WT, wild type; mut, mutation; FS, frameshift. (B) Representative sequence of the wild type (+/+) or BSG mutant clones generated with the different gRNAs. (C) Representative bright-field images of WT and BSG KO kidney organoids. Scale bars, 250 μm. Hematoxylin and eosin staining show tubular-like (^∗∗) and glomerular-like (^∗) structures. Scale bars, 250 and 50 μm (magnified views). (D) mRNA expression levels of BSG in WT and BSG KO kidney organoids by qPCR. Data are mean ± SD. n = 1 independent experiment from a pool of 12 organoids/group with at least two technical replicates each. (E) Experimental scheme for the infection with SARS-CoV-2 of WT and BSG KO kidney organoids. (F) TEM analysis of WT and BSG KO kidney organoids infected with SARS-CoV-2 at 3 dpi. Representative images of infected BSG WT specimens (left) show numerous viral particles in the cell surface of a dying cell (1 and 2). Details for podocyte-like cells (3) exhibiting podocyte-related structures including primary (pp) and the deposition of a basement membrane (bm). Scale bars, 2 μm, 200 nm (magnified views in 1 and 2), and 2 μm (magnified views in 3). Representative images of infected BSG KO specimens (right) show numerous viral particles in the intercellular space (4) and in the cell surface (5). Details for podocyte-like cells (6) exhibiting podocyte-related structures including primary (pp) and the deposition of a basement membrane (bm). Scale bars, 2 and 5 μm, 200 nm (magnified views in 4 and 5), and 1 μm (magnified view in 6). (G) SARS-CoV-2 mRNA expression levels in mock-treated or SARS-CoV-2-infected WT and BSG KO kidney organoids at 3 dpi by qPCR. Data are mean ± SD. n = 1 independent experiment from a pool of 12 organoids/group with at least two technical replicates each. (H) Immunofluorescence of mock or SARS-CoV-2-infected WT and BSG KO kidney organoids at 3 dpi for ACE2 (green), viral nuclear protein (NP, red), LTL (gray), and DAPI (blue). Scale bars, 250 μm. See also Figure S8.

Figure 6

ACE2 expression in diabetic human kidney organoids sustains SARS-CoV-2 infections (A) Experimental scheme for the infection with SARS-CoV-2 of WT and ACE2 KO kidney organoids under control or diabetic conditions. (B) Immunofluorescence of mock or SARS-CoV-2-infected specimens under control or diabetic conditions at 1 dpi for ACE2 (green), virus nuclear protein (NP, red), LTL (magenta), and DAPI (blue). Scale bars, 250 μm. n = 3 organoids per condition. (C) SARS-CoV-2 mRNA expression levels of mock or SARS-CoV-2-infected WT and ACE2 KO kidney organoids in control or diabetic conditions by qPCR. Data are mean ± SD. n = 1 independent experiment from a pool of 12 organoids/group with two technical replicates each. (D) ACE2 mRNA expression levels of mock or SARS-CoV-2-infected WT and ACE2 KO kidney organoids in control or diabetic conditions by qPCR. Data are mean ± SD. n = 1 independent experiment from a pool of 12 organoids/group with two technical replicates each. (E) Protein levels of ACE2 in WT and ACE2 KO kidney organoids under control or diabetic conditions by western blot analysis. β-actin was used as loading control. Data from a pool of 12 organoids/group are shown. (F) Lentiviral transduction of ACE2 in ACE2 KO kidney organoids (ACE2t) under control or diabetic conditions. Immunofluorescence of mock or SARS-CoV-2-infected specimens at 1 dpi for the viral nuclear protein (NP, red), ACE2 (green), LTL (white), and DAPI (blue). Scale bars, 250 and 50 μm (magnified views). n = 1 organoid (mock) and n = 2 organoids (SARS-CoV-2 infected). (G) Quantification of ACE2 expression (shown as integrated density-IntD) in (F). Data are mean ± SD. n = 2 organoid per condition performing two technical replicates. No statistically significant differences were observed. One-way ANOVA, Tukey’s multiple comparisons test. (H) Quantification of the area of NP⁺ cells in (F). Data are mean ± SD. n = 2 organoid per condition performing two technical replicates. ^∗p < 0.05, one-way ANOVA, Tukey’s multiple comparisons test. See also Data S1.

Figure 7

SARS-CoV-2 infection in tubular epithelial cells derived from diabetic human kidney biopsies (A) Representative bright-field images of HPTCs from non-diabetic (control) or diabetic patient kidney biopsies. Scale bars, 100 μm. (B) Seahorse analysis of control and diabetic HPTCs. The oxygen consumption rate (OCR) data are normalized to total protein. Basal respiration, cellular ATP production, and maximal respiration are shown. Data are mean ± SD from at least n = 3 biological replicates/group. ^∗p < 0.05; ^∗∗p < 0.005; ^∗∗∗p < 0.0005; ^∗∗∗∗p < 0.0001, one-way ANOVA, Tukey’s multiple comparisons test. (C) Experimental scheme for SARS-CoV-2 infection in control or diabetic HPTCs. Immunofluorescence of infected control or diabetic HPTCs at 1 dpi for the viral nuclear protein (NP, red) and DAPI (blue). Scale bars, 100 μm. (D) Quantification of NP⁺ cells in (C). Data are mean ± SD. n = 2 independent biological replicates per condition performing at least six technical replicates. ^∗∗∗∗p < 0.0001, one-way ANOVA, Tukey’s multiple comparisons test. (E) qPCR analysis of SARS-CoV-2-infected control or diabetic HPTCs at 1 dpi for the detection of SARS-CoV-2 mRNA. Data are mean ± SD. n = 2 independent biological replicates per condition with at least three technical replicates. ^∗∗p < 0.01; ^∗∗∗∗p < 0.0001, one-way ANOVA, Tukey’s multiple comparisons test. (F) Experimental scheme for SARS-CoV-2 infection in control or diabetic HPTCs treated with DCA or vehicle for 16 h prior infection. (G) qPCR analysis of mock or SARS-CoV-2-infected control or diabetic HPTCs exposed to DCA or vehicle at 1 dpi. Data are mean ± SD. n = 1 independent experiment with at least two technical replicates. ^∗∗p < 0.005; ^∗∗∗p < 0.001; ^∗∗∗∗p < 0.0001, one-way ANOVA, Tukey’s multiple comparisons test. See also Figure S9.